Work

Designing user engagement with text-based chatbots

Year: 2017
Type: dissertation
Study design: Clinical Trial
Abstract: A chatbot is a computer program that interacts with humans via auditory or textual methods and is increasingly being adopted to various everyday practices.Although the use of chatbots is increasing, there is little known about the specific chatbot characteristics that influence its user engagement.The aim of this study was to assess how different chatbot characteristics influence the user's perceptions towards chatbots and, ultimately, relate to user engagement.There were three main goals.The first goal was to assess if different types of response latency are associated with higher levels of perceived humanness.The second goal was to assess if higher levels of social presence can be associated with the inclusion of small talk to a conversation.The third goal was to investigate if the perceived humanness and social presence of chatbot mediate user engagement.For this experiment four chatbots were created using the Facebook Messenger platform.Each chatbot was randomly assigned to the participant.Overall, 196 participants took part in this online experiment.In all conditions, participants answered one Perceived Humanness question and completed the Social Presence questionnaire and the User Engagement Scale.Results show that higher levels of social presence and perceived humanness are associated with higher levels of user engagement with chatbots.Incorporating small talk in a conversation was associated with higher levels of social presence.No significant differences could be obtained between the groups with different response latencies.This research suggests that in order to achieve a higher user engagement with chatbots, chatbot developers should focus more on the creation of a socially oriented communication style, instead of making the chatbot appear as a human-like entity.
Author: Katkute, Kristina
Institution: Leiden University
Language: English
Subfield: Marketing
Open Access status: green
OpenAlex ID: W7027587484
Cites: 28
Cited by: 1
Related to: 10